Extract emails, phones, and URLs from a PDF — regex extraction in your browser, no AI credits needed
Contact information scattered across a PDF — email addresses, phone numbers, URLs — is one of the most common things people want to pull out programmatically. A vendor catalog, a conference attendee list, an academic paper's author block, an annual report's investor-relations section. The regex-based extractor surfaces all of them in seconds without AI cost. Here is what patterns it matches, the three workflows where regex extraction is exactly the right shape of tool, and the two cases where you should reach for the AI variant instead.
How the extractor finds contacts
The tool reads the PDF's text content (no AI involved), then runs regex patterns against the extracted text. Three patterns produce three columns of output:
- Email addresses. Standard RFC-5322 pattern (with reasonable simplifications for common-case email recognition). Matches alice@example.com, jane.doe+tag@subdomain.example.co.uk, etc. Doesn't match obfuscated emails like "jane (at) example (dot) com" — those need the paid AI variant.
- Phone numbers. Practical 10-15 digit patterns covering common international formats: +1-555-123-4567 (US), +91 98765 43210 (India), (020) 7946 0958 (UK), and variants. Heuristic filtering avoids false positives on long number sequences (page numbers, ISBN fragments, encoded IDs).
- URLs. http:// and https:// URLs. Matches whether the URL is hyperlinked in the PDF or appears as plain text. Doesn't match "example.com" without a protocol — those would need a heuristic that produces too many false positives on common phrases.
Three workflows where regex extraction is exactly right
Cases where you want the free, fast, in-browser tool:
- Vendor / customer list extraction. Catalogs, vendor lists, customer rosters. The contact info is structured and explicit; regex finds it reliably.
- Academic paper author block. Research papers list authors with email addresses for correspondence. Regex extracts every address in one pass.
- Annual reports / investor relations. Annual reports list investor-relations contacts, board email addresses, regional office phone numbers. All structured, all regex-friendly.
- Conference attendee lists. Many conferences publish attendee directories as PDFs. Extracting contact info for follow-up is a regex job.
- Government employee directories. Many ministries and PSUs publish staff directories. Phone numbers and email patterns are standardized within India-government contexts, regex-detectable.
Two cases where AI extraction beats regex
Cases where you should reach for AI Extract Contacts (paid) instead:
- Obfuscated contacts. Email addresses written as "jane (at) example (dot) com" or "jane[at]example[dot]com" to evade spam-scrapers. Phone numbers written as "call nine-eight-seven-six-five." Names written with custom separators. AI handles these; regex misses them entirely.
- Contacts embedded as images. Some PDFs deliberately show contact info as images rather than text to prevent automated harvesting. Regex sees no text; nothing to extract. AI variant runs OCR + visual extraction to surface the contacts despite the obfuscation.
Output formats and their use cases
Three output formats:
- Per-page table view (in-tool). Each contact shown with its source page. Useful for browsing and verifying the extraction before downloading.
- CSV download. Type / value / source page columns. Drop into a spreadsheet for further processing, deduplication, or import into a CRM.
- vCard download (for emails + phones). Standard vCard format for direct import into Apple Contacts, Google Contacts, Outlook. One-click address-book population from a directory PDF.
Privacy and ethics
Three considerations:
- Client-side extraction. Everything runs in your browser. The PDF never uploads. Important for documents that contain personal information about identifiable people — none of those bytes leave your machine.
- Cold-outreach legality. Just because you can extract someone's email doesn't mean cold-outreach is legal. Many jurisdictions (GDPR in EU, DPDP Act in India, CAN-SPAM in US) require legitimate basis for unsolicited contact. Extract for personal use, customer-relationship management, or legitimate-interest contexts; don't bulk-extract for spam.
- Source acknowledgment. If you extracted contact info from someone's published work (academic paper, annual report) and you reach out, mentioning the source is good professional practice. "I noticed your contact in the X annual report and wanted to reach out about Y." Sets the right expectation.
Limits and compatibility
On the free web tool, regex extraction handles PDFs up to 100 MB with no contact-count cap. Parsing runs in your browser; nothing is uploaded. Output is the extracted-contacts table plus CSV + vCard downloads.
Common pairings: AI · OCR first if the PDF is scanned (regex needs text to operate on). AI Extract Contacts when regex doesn't catch obfuscated patterns. Remove Metadata before sharing the extracted list to avoid leaking the source PDF's metadata into a recipient's hands.